304
So-inversion as Polarity Focus1
Jim Wood
New York University
1 Introduction
So-inversion, exemplified in (1a), is superficially similar to neither-inversion
(1b).
(1)
a.
b.
John plays guitar and so do I.
John doesn’t play guitar and neither do I.
In this paper, I propose that so-inversion is best analyzed as parallel to neitherinversion. Both constructions are examples of polarity-focus. So in so-inversion
is an affirmative polarity marker, accompanied by the focus particle too. However, too is not always pronounced.
1.1 Additivity
So-inversion clearly means something like ‘also’. It is only felicitous in circumstances where also, too, or as well are also felicitous. I will refer to this property
as ‘additivity’. Thus, (2a-b) are additive while (2c) is not. So-inversion is also
additive.
(2)
a.
b.
c.
I also play guitar.
I play guitar too.
I play guitar.
1.2 So and Polarity
So-inversion interacts with polarity in important ways. So-inversion generally
requires an affirmative antecedent (Klima 1964), unlike other additive
constructions (3c).
(3)
a.
b.
c.
*
*
John does not play guitar and so do I not (play guitar).
John doesn’t play guitar and so do I.
John does not play guitar and I also do not (play guitar).
305
So-inversion itself must be affirmative (4a), unlike other additive constructions
(4c).
(4)
a.
b.
c.
* John plays guitar but so don’t I.2
* John plays guitar but so do I not.
John plays guitar but I don’t <also> play guitar <also>.
Conversely, neither-inversion requires a negative antecedent.
(5)
a.
b.
*
John doesn’t play guitar and neither do I.
John plays guitar and neither do I (play guitar).
Neither itself must be negative, just as so must be affirmative.
1.3 So and too
As mentioned above, so-inversion clearly has additive meaning. The focus particle too also induces additive meaning (2b). But interestingly, too can be added
to so-inversion without any extra additive meaning.
(6)
a.
John plays guitar, but so too does Mary.
= John plays guitar, but so does Mary.
Just as some children ignore their parents, so too do some
parents ignore their children.
= Just as some children ignore their parents, so do some
parents ignore their children.
b.
Although they co-occur in the preposed position, additive-too and additive-so
are in an interesting kind of complementary distribution.
(7)
(8)
a.
b.
*
So do I.
Too do I.
a.
b.
*
I do too.
I do so.
(ungrammatical on relevant reading)
Additive-too cannot occur alone in the preposed position, unlike so (7).
Additive-so can only occur in the preposed position (8).
1.4 So and N-either
I claim that so-inversion and neither-inversion have strongly similar syntactic
derivations. Huddleston & Pullum (2002:1539) claim that there are two crucial
differences between so-inversion and neither-inversion. The first is that so must
306
contrast subjects, but neither need not do so. This seems to be incorrect: so does
not have to contrast the subject.3 Example (9d) comes from President Barack
Obama’s nomination acceptance speech at the Democratic National Convention
in 2008.
Their claim is on the difference between (10a-b) and (10c-d).
However, the examples in (10) are much, if not completely, improved if and is
changed to but and too is added.
Thus, this does not seem to be a clear difference between so-inversion and
neither-inversion.
The second claimed difference is that so can occur with too, but neither stands
alone. One cannot, for example, insert not/nor/either with neither-inversion.
I would like to claim that this difference stems from a decomposition of neither
in to at least n- + either.
307
The reason neither cannot co-occur with a focus particle, then, is that it is n- that
always occurs with a focus particle, namely, either. This is analogous to
analyzing so-inversion as involving so + TOO.
2 What do we have to account for?
Any analysis of so-inversion should explain: (a) why it induces subject-aux inversion, (b) its obligatory affirmative polarity, and antecedent polarity matching,
(c) its co-occurrence with too, (d) the similarities with neither-inversion, and (e)
its additive meaning. In the next section, I show how my analysis accounts for
these properties.
2.1 Polarity Focus Constructions (PFCs)
308
A parametric difference between various analyses of is (a) is in complementary distribution with polarity markers, as in Russian, Basque, Finnish, (and
English), and (b) is realized by a separate polarity focus particle, and is not in
complementary distribution with polarity markers, as in Dutch, Hungarian, Oevdalian. A possible account of this variation, which I do not pursue in detail here,
would to take the first set to involve movement of PolP (AffP/NegP) to P to
focus polarity, and the second set to involve P being realized by a separate
lexical item. So-inversion falls under the first category.
2.2 So-inversion as Polarity Focus
The general framework assumed here is the Minimalist Program (Chomsky
1995, 2008, Collins 1997), assuming Antisymmetry (Kayne 1994) and no covert
movement (Kayne 1998).
Too, as a focus particle, is a focus head which merges in the left periphery of
the extended VP.5 This is in line with Kayne’s (1998) analysis of focus particles, including too. It attracts the focused constituent to its specifier (14).
Next, the affirmative polarity particle so merges with FocP, creating AffP.6 The
derivation proceeds normally from here, until merges with TP and attracts
PolP to its specifier after VP has evacuated it (15).7
309
Now consider neither-inversion. The key to accounting for the differences between so and neither is the decomposition of neither into n-+ either. The one
‘word’ neither does the work of the two words so too (16).8 The rest proceeds
the same (17).
2.3 Polarity Focus and the Facts of So-inversion
How does this account for the properties of so-inversion discussed above?
Subject-aux inversion is a general property of polarity focus. Whatever the
account of (18) is, this will extend to so-/neither-inversion.9
(18)
Never have I seen such a hideously ugly car.
Polarity restrictions follow directly. If so is an affirmative polarity particle, it
should not be compatible with negative polarity. Additive meaning comes from
the (sometimes silent) focus particle too, or (the always present) either. The latter cannot be silent because it forms a prosodic word with n-. Too and either are
focus particles, and are tightly connected with polarity focus. This connection
seems cross-linguistically ubiquitous. Finally, so-inversion in this analysis is di-
310
rectly analogous to neither-inversion. The differences stem from decomposition
of neither into n-+ either.
3 Previous Analyses
I now consider some previous analysis of so-inversion, and show that they fail to
capture the properties outlined above. Further, they make incorrect predictions
of their own.
3.1 So as a pro-form
In a recent analysis Toda (2007) argues that so is a pro-form that replaces a
preposed VP. This is very similar to the analysis of so-inversion in Chomsky
(1957:65-6). This has intuitive appeal, since VP-ellipsis is very common with
so-inversion: if so replaces the VP, this is explained.
However, such an analysis has several drawbacks. First, subject-aux inversion
must be stipulated. Second, there is no reason to expect a polarity restriction.
Third, there is no obvious relationship with too or neither. Fourth, the additive
meaning of the construction actually has to be denied (Toda 2007:fn6). While
Toda’s analysis gets many facts right for a few examples, it doesn’t extend,
within the language, beyond the construction itself.
I will not criticize Toda’s proposal directly, but wish to point out a more general problem with any “so as proform” approach: the verb phrase material still
seems to be there, either overtly or as ellipsis. This is most clearly shown in the
sentences like (6b) and (9).
Even when ellided, though, the VP can still be shown to be present. Compare
the behavior of so-inversion with a more clear verbal pro-form do so.10 In soinversion, wh-echo can be recovered, unlike with do so.
Similarly, so-inversion can co-occur with a Hanging Topic (Cinque 1977): 11
311
If so were a pro-form, we would not expect (19a-ii), (19a-iii), or (20a) to be
grammatical, since so would replace the constituent containing what in (19a-ii),
who in (19a-iii), and bugs in (20a).
3.2 So as an adjunct
So is taken to be a “connective” adjunct in Huddleston & Pullum (2002) and an
“additive adverb” in Quirk et al. (1985). I discuss these in turn.
A connective adjunct establishes a relationship between two sentences. Examples include moreover, nevertheless, etc. However, most connective adjuncts do
not induce subject-aux inversion (21b), nor do they exhibit polarity restrictions
(21c). The relationship between so-inversion and too and neither could be explained by calling them connective adjuncts too, but *Too do I would not be
explained. Additive meaning would seem to be partially explained on this hypothesis.12
Further, most connective adjuncts can appear in multiple places throughout the
phrase, unlike so in so-inversion.
An additive adjunct is an adverbial which is additive in the sense defined above.
Examples include also, as well, and too. This is closer to the present analysis,
since here there is an additive element present (i.e. too). However, most additive
adjuncts do not induce subject-aux inversion (23).13 Polarity restrictions also
would not be expected (24). The relationship with so and too might be expected,
though the co-occurrence of so too might be expected to be odd.
312
Thus, neither the adjunct nor pro-form approach to so-inversion capture the facts
of the construction.
4 New England so don’t I
In most of Eastern New England, there exists a construction usually referred to
as the so don’t I construction.14 It is important to note that so don’t I is affirmative. Some naturally occurring examples from Google are shown below.15
No similar construction exists for the negative (26a), it is obligatorily affirmative and cannot take a negative antecedent (26b), and there is no noninverted version (26c). The n’t does not license NPIs (27).
313
Unlike standard so-inversion, so don’t I does not allow an optional too (28). In
addition, for no speaker is an unreduced negative not allowed (29).
(28)
a.
b.
c.
d.
(29)
a.
b.
He plays guitar, but so (*too) don’t I.
He plays guitar, but so too do I.
Just as some parents ignore their children, so (*too) don’t some
children ignore their parents.
Just as some parents ignore their children, so (too) do some
children ignore their parents.
* He plays guitar, but so do not I.
* He plays guitar, but so do I not.
The pragmatic force of so don’t I is one of implicature canceling. That is, (30a)
is only pragmatically felicitous when there exists an implicature like (30b).
Why is (26a) ungrammatical? While I do not have the space to go into the details, I would like to argue that this implicature cancelation is syntactically
represented by an abstract negation morpheme, which constituent negates the
verb phrase. Thus, so don’t I is similar to double negation (31b), which is
possible in ellipsis contexts (32).
If this is on the right track, then the impossibility of neither can’t I in New
England English would relate to the impossibility of neither as constituent
negation in double negation sentences (33).
Since neither must move to the clausal negation position (16), bypassing not, we
predict that not can be the constituent negator, and neither the clause negator,
but this would preclude n’t. This prediction is borne out:
314
The ungrammaticality of (33) and (26a) are then related: neither must be the
clausal negator, at the expense of n’t. In so don’t I, the lower polarity operator so
is not negative, allowing merger of n’t but with scope only over the lower implicature similar to metalinguistic negation (Carston 1996).
5 Conclusion
So-inversion involves polarity focus, where so is best understood as an affirmative polarity particle. The properties of so-inversion, which extensively
differentiate it from pro-forms and adjuncts, are very similar to neither-inversion
specifically, and polarity focus in general. The differences between so and
neither can be shown to follow from the fact that neither is composed of (at
least) two morphemes, n- and either. Dividing the functions of polarity and
additivity into two separate elements allows for a clearer understanding microparametric variation in New England so don’t I. Finally, analyzing so-inversion
as polarity focus invites interesting connections with polarity-focus constructions cross-linguistically, which very often involve additive particles like
too or also, as well as a high polarity focus position.
Notes
1
I am utterly indebted to the following people for inspiring discussion of the material here: Mark
Baltin, Inna Livitz, Jeroen van Craenenbroeck, Richard Kayne, Tricia Irwin, Oksana Laleko and
Sara Schmelzer, as well as others I can’t think of at the moment. In addition, a huge thanks to
Melinda Kaye Wilson for extensive discussion of the so don’t I data.
2
Though see below on so don’t I.
3
Examples (9a-c) are from the Contemporary Corpus of American English (COCA; Davies 2008).
Many more can be found.
4
The object of the preposition for in (b), this divorced father, appears to be the speaker, a
phenomenon discussed in detail by Collins & Postal (2008).
5
I assume more VP structure than is shown here. See Alexiadou & Anagnostopoulou (2004), Collins
(2005), and Baltin (to appear).
6
More generally, PolP; see Culicover 1991. Alternatively, there may be reasons to believe that so is
in Spec,PolP. See Haddican (2004).
7
Here I show VP moving to SpecZP, the latter adopted from Baltin (2006). See also Kayne (2005)
and Bentzen (2005) for the possibility of a position between T and Pol to which VP moves.
8
Here I show this as a result of movement – the idea is that negative constituents like neither must
move to NegP to establish clausal negation (see Kayne 1998, den Dikken 2006). Another possibility
is that nothing moves and n- realizes Neg.
9
See Haegeman (2000) for discussion. It is not clear how this would extend to the system in Sobin
(2003).
315
10
See Baltin (to appear) and Haddican (2007) on do so as a pro-form. It is worth noting that Toda
(2007:fn3) explicitly argues that his pro-form so has nothing to do with the pro-form do so. This
doesn’t affect the arguments here, though: pro-forms usually don’t occur with the constituents they
supposedly replace, and in Toda’s analysis, the VP is swallowed up whole, derivationally, by the
pro-formalization process.
11
The important thing here is the contrast. Thank you to van Craenenbroek (p.c.) for pointing out
some reasons to believe that this is something like Hanging Topic, rather than Topicalization or Left
Dislocation.
12
That is, connective adjuncts seem to be additive in the clausal sense. For example, moreover could
be replaced by also in many cases. Usually, though, so focuses a specific sub-constituent of the
clause, such as the subject, and not the clause itself. It seems much harder to get moreover to do this.
13
An exception is the very interesting case of as do I and nor do I, which also arguably involve
polarity in some important way. See Potts (2002) for some discussion.
14
See Lawler (1974), Labov (1972), Pappas (2004), Freeman (2004), Horn (1978, 2008), and Gilman
(1989).
15
These examples are from the following websites, in order:
http://www.yelp.com/biz/runway-new-york-3
http://stampinangeljenn.blogspot.com/2008/06/girls-daynight.html
http://www.sunjournal.com/story/258000-3/LewistonAuburn/Students_grades_to_go_online/
http://www.boston.com/news/politics/politicalintelligence/2008/07/mccain_sharpens.html
16
In some dialects, sentences of this kind are acceptable and negative. Speakers who judge these as
acceptable are not hard to find, nor are examples on the internet. Unfortunately, I do not have time to
discuss such speakers here. Importantly, they are ungrammatical for speakers of the dialect in
question. All of my informants unequivocally rejected such sentences, often asking what such an
expression could possibly mean.
References
Alexiadou, Artemis and Elena Anagnostopoulou. 2004. “Voice morphology in
the causative-inchoative Alternation: evidence for a non unified structural
analysis of unaccusatives”. In The unaccusativity puzzle: explorations of the
syntax-lexicon interface, edited by Artemis Alexiadou, Elena Anagnostopoulou
and Martin Everaert, 115-136. Oxford: Oxford University Press.
Baltin, Mark. 2006. “The Nonunity of VP-Preposing”. Language 82 (4):734766.
Baltin, Mark. To appear. “Deletion Versus Pro-Forms: An Overly Simple
Dichotomy?” Natural Language and Linguistic Theory.
Bentzen, Kristine. 2005. “What’s the better move? On verb placement in
Standard and Northern Norwegian”. Nordic Journal of Linguistics 28 (2):153188.
Carston, Robyn. 1996. “Metalinguistic negation and echoic use”. Journal of
Pragmatics 25 (3):309-330.
Chomsky, Noam. 1957. Syntactic Structures. Mouton: The Hague.
Chomsky, Noam. 1995. The Minimalist Program. Malden, MA: MIT Press.
Chomsky, Noam. 2008. “On Phases”. In Foundational Issues in Linguistic
Theory: Essays in Honor of Jean-Roger Vergnaud, edited by Robert Freidin,
Carlos P. Otero and Maria Luisa Zubizarreta, Cambridge, MA: MIT Press.
Cinque, Guglielmo. 1977. “The Movement Nature of Left Dislocation”.
Linguistic Inquiry 8:397-412.
316
Collins, Chris. 1997. Local Economy. Cambridge, MA: MIT Press.
Collins, Chris. 2005. “A smuggling approach to the Passive in English”. Syntax
8 (2):81-120.
Collins, Chris and Paul Postal. 2008. Imposters. Manuscript, New York
University. Available on lingBuzz/000640.
van Craenenbroek, Jeroen. 2004. Ellipsis in Dutch dialects. Doctoral
Dissertation, Leiden University.
Culicover, Peter. 1991. “Polarity, inversion, and focus in English”. ESCOL 8
(1):46-68.
Davies, Mark. 2008. The Corpus of Contemporary American English (COCA):
385 million words, 1990-present. Available online at
http://www.americancorpus.org.
den Dikken, Marcel. 2006. “Either-float and the syntax of co-or-dination”.
Natural Language and Linguistic Theory 24 (3):689-749.
Freeman, Jason. 2004. “Syntactic Analysis of the “So Don’t I” Construction”. In
Cranberry Linguistics 2: University of Connecticut Working Papers in
Linguistics, edited by Miguel Rodríuez-Mondoñedo and M. Emma Ticio, 2538.
Gilman, E. Ward (ed). 1989. Mirriam-Webster Dictionary of English Usage.
Springfield, MA: Merriam-Webster.
Haddican, Bill. 2004. “Affirmation and weak islands”. Snippets 9:9-10.
Haddican, Bill. 2007. “The Structural Deficiency of Verbal Pro-Forms”.
Linguistic Inquiry 38:539-547.
Haegeman, Liliane. 2000. “Inversion, non-adjacent inversion and adjuncts in
CP”. Transactions of the Philological Society 98 (1):121-160.
Holmberg, Anders. 2001. “The syntax of yes and no in Finnish”. Studia
Linguistica 55 (2):140-174.
Holmberg, Anders. 2007. “Null subjects and polarity focus”. Studia Linguistica
61 (3):212-236.
Horn, Laurence R. 1978. “Some aspects of negation”. In Universals of Human
Language, Vol. 4: Syntax, edited by Greenberg et al. 127-210. Stanford:
Stanford University Press.
Horn, Laurence R. 2008. Postface for NHN3. In A Natural History of Negation
[Japanese Translation], edited by Laurence R. Horn.
Huddleston, Rodney and Geoffrey K Pullum. 2002. The Cambridge Grammar of
the English Language. Cambridge: Cambridge University Press.
Johnson, Kyle. 2001. “What VP Ellipsis Can Do, and What it Can’t, but not
Why”. In The Handbook of Contemporary Syntactic Theory, edited by Chris
Collins and Mark Baltin, 439-479. Malden, MA: Blackwell.
Kayne, Richard. 1994. The Antisymmetry of Syntax. Malden, MA: MIT Press.
Kayne, Richard S. 1998. “Overt vs. Covert Movement”. Syntax 1 (2):128-191.
Kayne, Richard S. 2005. Movement and Silence. New York: Oxford University
Press.
Kayne, Richard S. 2006. “On Parameters and on Principles of Pronunciation”. In
Organizing Grammar: Linguistic Studies in Honor of Henk van Riemsdijk,
edited by H. Broekhuis, N. Corver, R. Huybregts, U. Kleinhenz and J. Koster,
Berlin: Mouton de Gruyter.
Klima, Edward. 1964. “Negation in English”. In The Structure of Language,
317
edited by Jerry Fodor and Jerrold Katz, 246-323. Englewood Cliffs, NJ:
Prentice-Hall.
Labov, William. 1972. “Negative attraction and negative concord in English
Grammar”. Language 48 (4):773-818.
Laka, M.A. Murgaza. 1990. Negation in Syntax: On the Nature of Functional
Categories and Projections. Doctoral Dissertation, MIT.
Laleko, Oksana. 2006. NET-licensed VPE in Russian: An InformationStructural Account of a Syntactic Puzzle. Paper presented at the LSA Annual
Meeting.
Lawler, John. 1974. “Ample Negatives”. Chicago Linguistics Society 10:1-24.
Lipták, Anikó. 2003. The expression of sentential emphasis in Hungarian.
Manuscript, Leiden University.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik. 1985.
A comprehensive grammar of the English language. London: Longman.
Pappas, Dino Angelo. 2004. “A Sociolinguistic and Historical Investigation of
“So don’t I””. In Cranberry Linguistics 2: University of Connecticut Working
Papers in Linguistics, edited by Miguel Rodríuez-Mondoñedo and M. Emma
Ticio, 53-62.
Potts, Chris. 2002. “The lexical semantics of parenthetical-as and appositivewhich”. Syntax 5 (1):55-88.
Rizzi, Luigi. 1997. “On the fine structure of the left periphery”. In Elements of
Grammar – Handbook of Generative Syntax, edited by Liliane. Haegeman,
281-338. Kluwer: Dordrecht.
Rosenkvist, Henrik. 2007. “Subject Doubling in Oevdalian”. Working Papers in
Scandinavian Syntax 80:77-102.
Sobin, Nicholas. 2003. “Negative Inversion as nonmovement”. Syntax 6 (2):183-212.
Toda, Tatsuhiko. 2007. “So-inversion re-visited”. Linguistic Inquiry 38:188-195.
Jim Wood
New York University
Department of Linguistics
726 Broadway, 7th Floor
New York, NY 10003
[email protected]